← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
4k
tokens
3,913
PP tok/s
128.3
TG tok/s
1047
TTFT (ms)
20.0
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.7
macOS
macOS 26.4.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,290 | 134.8 | 19.3 GB | view |
| 4k | 3,913 | 128.3 | 20.0 GB | current |
| 8k | 4,176 | 123.0 | 20.4 GB | view |
| 16k | 4,035 | 115.4 | 21.0 GB | view |
| 32k | 3,433 | 105.8 | 22.4 GB | view |
| 64k | 2,457 | 88.1 | 25.0 GB | view |
| 128k | 1,528 | 66.4 | 30.4 GB | view |
| 195k | 940.5 | 46.4 | 36.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 134.8 | 1.00× |
| 2× | 213.4 | 1.58× |
| 4× | 299.8 | 2.22× |